Whisper Small kh - Sethisak San
This model is a fine-tuned version of openai/whisper-small on the KhmerAsrDataset dataset. It achieves the following results on the evaluation set:
- Loss: 0.1773
- Wer: 96.2162
- Cer: 26.6836
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 4000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 0.0916 | 3.2154 | 1000 | 0.1611 | 97.4775 | 30.2389 |
| 0.0112 | 6.4309 | 2000 | 0.1518 | 96.3964 | 27.3481 |
| 0.0015 | 9.6463 | 3000 | 0.1706 | 97.2973 | 26.4233 |
| 0.0003 | 12.8617 | 4000 | 0.1773 | 96.2162 | 26.6836 |
Framework versions
- Transformers 4.48.0
- Pytorch 2.8.0+cu126
- Datasets 2.14.7
- Tokenizers 0.21.4